Annual Journal of Hydraulic Engineering, JSCE, Vol.54, 2010, February
Annual Journal of Hydraulic Engineering, JSCE, Vol.54, 2010, February
GEOMETRY EFFECT ON FLOW AND SEDIMENT
DEPOSITION PATTERNS IN SHALLOW BASINS
Sameh A. KANTOUSH1, Tetsuya SUMI2 and Anton J. SCHLEISS3
1Member of JSCE, Dr. of Eng., Postdoctoral researcher, Water Resources Research Center, Disaster Prevention
Research Institute, Kyoto University (Goka-sho, Uji-shi, 611-0011, Japan)
2Member of JSCE, Dr. of Eng., Professor, Water Resources Research Center, Disaster Prevention Research Institute,
Kyoto University (Goka-sho, Uji-shi, 611-0011, Japan)
3Professor, Director, Ecole Polytechnique Fédérale de Lausanne (EPFL), Laboratory of Hydraulic Constructions
(LCH) (Station 18, CH -1015 Lausanne, Switzerland)
The design of stilling basin geometry of flood mitigation dams in view of clear water flow may
differ when considering of the water-sediment and deposition effects. To ensure effectiveness of energy
dissipation, appropriate design of stilling basin leading to optimal geometry is required. The present study
investigates the influence of the shallow basin geometry on flow and sediment deposition patterns. The
effect of the geometry was investigated using systematic physical experiments, numerical simulation, and
field data. This allowed to identify the optimal stilling basin shape of the flood mitigation dam that
dissipates more energy, minimizes the deposition and reduces the number of horizontal recirculation cells.
The results help to understand the influence of geometry on the flow stability, sediment deposition and
flushing process in order to choose the appropriate stilling basin geometry. The prediction of sediment
profile depends on the prediction of flow pattern and, in turn, sediment deposits were able to change the
flow structure.
Key Words: Influence of geometry, flood retention dam, flow pattern, sedimentation pattern, stilling basin
1. INTRODUCTION
Flow separation and reattachment due to sudden
changes in geometry in internal flow occur in many
engineering applications such as in open channels,
shallow reservoirs, groin fields and stilling basins.
The occurrence of large-scale instabilities and
consequent vortex formation in planar jets in
various geometries has been studied1). In shallow
flow both the limited depth and the bottom friction
influence the development of the mixing layer2). The
limited depth restricts the large structures in the
mixing layer to basically 2D horizontal motions.
A model was developed for obtaining the optimal
dimensions of stilling basin and its appurtenances3).
In symmetrical geometries of stilling basins for a
fixed and horizontal channel bed, and various
expansions and aspect ratios, the hydraulic jumps
have been classified4). The hydraulic jump moves
upstream and its front gradually become normal
(R-jump) as the tailwater level is increased4).
Several ratios of the stilling basin width to the inlet
channel width were analyzed, which affect flow
stability, necessary tailwater depth, and length for
the hydraulic jump5). A more detailed analysis of
hydraulic jumps in abruptly expanding channels was
provided6). Many experimenters have observed that
the flow in symmetric expansions is asymmetric.
The asymmetric flow in sudden expanding stilling
basins may cause asymmetric scour downstream the
basin.
Numerical and theoretical stability analysis of
shallow flows in a series of rectangular reservoirs
compared with experimental data has been
discussed7). A series of numerical simulations and
compared with scaled laboratory experiments, to
investigate the sensitivity of flow and sediment
parameters were presented8).
2. OBJECTIVES AND TARGET PROJECT
- 133 -
Design of stilling basin of flood mitigation dams
Mixing tank
Measured sediment on December 2006
Sediment supply tank
Mixer
Drainage
(b)
(a)
Pipe
Water supply
Valve
z
Stilling basin
Flow meter
Distributor
y
Movable frame
UVP
Video camera
PVC wall
Slits for
flushing
B = 4.0 m
Flap gate
Picture was taken on December 2006
To settling basin
Fig.2 Schematic view of experimental facility installation.
Measured sediment volume is 1,036 m3
Fig.1 Stilling basin of Masudagwa dam (a) schematic picture of
flow and deposition; (b) measured sediment profile and volume
such as Masudagawa dam in Shimane Prefecture is
governed by several parameters such as: number of
bottom outlets, approach Froude number, basin
geometry, and tailwater level. Flood mitigation dam
is a gateless outlet dam designed only for the
purpose of flood control which provides long-term
and efficient protection against floods and is called
dry dam in USA.
The study focuses on the use of the findings of
the fundamental research study about influence of
geometry to define the stilling basin in flood
mitigation dam. Case of study for stilling basin of
Masudagawa dam, newly completed flood
mitigation dam in Japan is shown in Fig. 1. The
stilling basin of Masudagawa dam dissipates the
energy by hydraulic jump type equipped with an
endsill and slits for flushing sediment and allowing
fish passage. The flow width from the bottom outlet
shows a narrow flow width comparing with the
width of the downstream river (Fig.1(a)).
Horizontal recirculation cells are formed at the
sides of the inflow jet. These recirculation cells
combined with sediment transport are harmful.
During the first and last stages of a flood when the
reservoir water level is reduced, a huge sediment
outflow occurs. While increasing the basins water
level, the velocity is increased and large sediment
volumes are deposited in the basin as shown in
Fig.1(b). Therefore, it is possible to minimize the
damage by stopping the horizontal recirculation
cells in the stilling basin during the high water level
conditions. An asymmetric flow pattern with a low
number of circulation cells is favorable to reduce the
exchange processes between flow, bed deposition
and the damage of basin slits. Flow pattern with
large stagnant zones allows only a portion of flow to
pass the basin in a time period less than the settling
time. Beside the purpose of controlling the number
of recirculation cells and deposition, the objective is
also to gain deeper insight into the physical
processes of sedimentation in shallow basins. The
findings will help designers of stilling basins of new
flood mitigation dams to predict the flow and
sediment deposition and to choose an appropriate
geometry. To obtain an optimum design and
operation of bottom outlets and stilling basin, 3D
numerical modeling will be carried out for different
scenarios. The main criterion for the different
simulation is to find out which geometry dissipates
more energy and transports the maximum volume of
sediment over time.
3. METHODOLOGY
The complexities associated with predicting the
flow and sedimentation pattern in various basin
geometries emphasis the need of combination of the
numerical results with the physical experiments.
This helped to deeply understand the physics and
the process of the sedimentation in the shallow
basins.
Sixteen experiments with clear water and
water-sediment mixture were performed in a facility
(Fig.2) at the Laboratory of Hydraulic Constructions
(LCH) of the Swiss Federal Institute of Technology
(EPFL). The experiments have been conducted in a
rectangular shallow basin shown in Fig. 1(b), with
inner maximum dimensions of 6 m in length and 4
m in width. The inlet and outlet rectangular
channels are both 0.25 m wide and 1 m long. The
influence of different shallow reservoir geometries,
with variable widths and lengths has been achieved
experimentally. To investigate the effect of basin
width on the flow and sedimentation processes, first
the 6 m long basin was performed (Fig. 1(a)).
The width was then reduced successively from 4
to 3 to 2, to 1, and to 0.50 m. In a second step the
effect of basin length was examined by reducing the
length of the rectangular shallow basin from 6 m to
5, to 4, and to 3 m successively.
- 134 -
(a)
(b)
0.135
0.123
0.111
0.099
0.087
0.074
0.062
0.050
0.038
0.026
0.014
0 m/s
0.135
0.123
0.111
0.099
0.087
0.074
0.062
0.050
0.038
0.026
0.014
0 m/s
Fig.3 Time averaged flow pattern and velocity
magnitude (a) clear water; (b) sediment
entrainment flow.
Fig.4 Long-term morphological evolution of deposition at 4 runs with
Q =7.0 l/s, h = 0.2 m and sediment concentration C = 3.0 g/l.
To model suspended sediment flow in the
laboratory model, crushed walnut shells with a
median grain size d50=50 μm, and a density of 1500
kg/m3 were used in all tests. Several parameters
were measured during every test, namely: 2D
surface velocities, thickness of the deposited
sediments, as well as concentration.
Two dimensional depth-averaged flow and
sediment transport simulations representing the
experimental set-up have been performed with the
numerical model CCHE2D9), developed by NCCHE
(National Center for Computational Hydroscience
and Engineering) based on a variant of the finite
element method. The model was used to predict
river flow patterns and related bed and bank erosion
for both uniform and non-uniform sediment
transport. Both depth-averaged k–ε and eddy
viscosity turbulence closures are available.
CCHE2D was used for its capabilities to simulate
suspended sediment transport.
4. RESULTS AND DISCUSSIONS
(1) Flow pattern with and without sediment
The flow features and large–scale structures
were investigated by using LSPIV measurement
technique. Fig. 3(a) shows an overview of the
velocity field and behavior of large-scale coherent
structures in clear water. A plane jet issues from the
narrow leading channel and enters straight ahead in
the first half meter of the wider basin. Then, the
main flow tends to develop in a curved way towards
the right hand side over the next two meters until it
touches and follows the right wall. When separating
from the right wall, the main flow induces a
recirculation zone. A main large stable circulation
cell is generated in the centre of the basin rotating
anticlockwise. Two small triangular cells are formed
rotating clockwise in the upstream corners of the
basin. The two corner cells are dependent from each
other and alternatively change. Both are controlling
the size of the main gyre. The jet preference for the
right side is weak, since a stable mirror image of the
flow pattern can easily be established by slightly
disturbing the initial conditions.
The stable asymmetric pattern, with a larger and
smaller recirculation zone at the right and left
corners, can be explained by a Coanda effect by
which any perturbation of the flow field, pushing
the main flow to one side of the basin, gives rise to
larger velocities at the left side along the jet. The
entrainment of mass into the jet creates a mass flux
bigger than the discharge, which contributes to the
recirculation. When the distance between the inflow
and outflow section is big enough, the two are
decoupled and the recirculation has room to push
the jet aside, which is helped by the Coanda effect.
Fig.3(b) shows the second flow feature
developed with sediment entrainment. The addition
of sediment decreases the mixing length or
increasing the eddy size of the right corner gyre
with time. The flow becomes more stable and
symmetric as shown in Fig. 3(b). As a result of
ripple formation (see Figs. 4(a), (b)) and suspended
sediment concentrations, the flow field is
completely changed as shown in Fig. 3(b). The
gyres in the upstream corners disappeared and a
pattern emerges rather symmetric with respect to the
center line (see flow pattern schema in Fig. 4). The
two remaining gyres interact with the jet which
shows tendency to meander. The changes in the bed
forms or effective roughness resulting from the
sediment deposition can completely modify the
overall flow pattern.
- 135 -
0.18
Y
X
0.14
X =1.5 m
0.12
B =2 m
0.10
18.0hr
0.06
0.04
9.0hr
0.02
4.5hr
3.0hr
1.5hr
0.00
‐2.00
‐1.50
‐1.00
‐0.50
0.00
0.50
1.00
1.50
Fig.6 Flow patterns, (left) and evolution of deposition, (right)
for reservoir width (a) B = 3.0 m, (b) B = 2 m, at 4.5 hrs
0.10
2.0
0.09
Water surface
Deposits after 1.5hr (T7)
Y
Deposits after 3.0hr (T7)
Deposits after 4.5hr (T7)
Y‐distance in transverse direction [m]
Bed thickness deposited [m]
0.08
Fig.5 Cross section bed profiles for reference geometry
As a conclusion, a strong interaction between
flow field and bottom topography occurred in the
basins inlet region with a quasi equilibrium state.
There, most of the bed form features disappeared, as
the deposition was increased fairly fast. Numbers of
recirculation cells that exist in the basin have a
strong influence on the flow and sediment
deposition behavior. By increasing the numbers of
cells, the deposited volume of sediments in the basin
increased. The bed morphology and the
corresponding schematic flow field for the reference
geometry are shown in Fig.4 for four different runs
(1.5, 3, 4.5, 9 hrs) allow a comparison of the
long-term bed evolution in the rectangular basin.
For all runs, two sedimentation behaviors were
observed. First is the development of the sediment
deposition developed with ripples formation
concentrated on the right hand side till bed thickness
deposition reaches up to 15% of the water depth as
shown in Fig.4 (a). Then the sediment deposition
concentrated along the centerline with relatively
steep gradients near the inlet channel and the first
part of the jet (Fig.4 (b)). Deposits configurations in
Fig. 3 shows how the mixture of water and sediment
is diffused throughout the basin following the
general flow patterns. The footprint of the flow
patterns was clearly visible in the morphology.
However, the increased roughness height associated
with mobile sediment may contribute to increase in
shear velocity. A symmetric ripples pattern forming
in the middle of the basin is clearly visible after 4.5
hrs (Fig. 4(c)). After 9.0 hrs (Fig. 4 (d)), the
deposition on the center gradually increased
generating a wider bed elevation underneath the jet
centerline.
Fig. 5 illustrates the depositions in transversal
direction of the basin at a distance of 1.5 m from the
inlet, for the 5 runs. After 1.5 hrs, almost uniform
depositions over the basin with average thickness of
0.015 m are observed. Due to the complete change
of the flow pattern after 3.0 hrs, sediment deposition
B = 3.0 m
0.08
T7
Deposits after 1.5hr (T8)
Deposits after 3.0hr (T8)
0.07
Deposts after 4.5hr (T8)
Y
T8
0.06
X
X =3.0 m
B = 2.0 m
Bed thickness deposited [m]
0.16
Deposits after 1.5hr
Deposits after 3.0hr
Deposits after 4.5hr
Deposits after 9.0hr
Deposits after 18.0hr
B=3m
0.20
0.05
X
0.04
0.03
4.5hr 3.0hr
0.02
0.01
1.5hr
0.00
-2.00 -1.75 -1.50 -1.25 -1.00 -0.75 -0.50 -0.25 0.00
Left bank
0.25
0.50
0.75
Y-distance in transverse direction [m]
1.00
1.25
1.50
1.75
2.00
Right bank
Fig.7 Bed thickness evolution of the reduced widths of B = 3.0,
2.0 m at 1.5, 3.0 and 4.5 hrs for cross section at X =3.0 m
rate is slightly increased by 0.005 m. Fig. 5 shows
almost constant sediment deposits within the first
hours for run 2, run 3, and run 4 but the deposits rate
is increased for runs 4 and 5. It may be concluded
that a stable morphology has been reached after 18.0
hours and almost morphological equilibrium in the
basin has been reached for. Bed forms are almost
uniform after 1.5 and 3.0 hours but 4.5 hr deposits
show wavy bed forms (Fig.5). The bed thickness
observed after 9.0 hrs is almost three times higher
than for 4.5 hrs. Almost 50% of the basin volume
was filled by the deposits after 18 hrs.
(2) Influence of the basin width
Flow patterns and associated bed depositions
after 4.5 hrs for reduced width basins of 3 and 2 m,
are shown in Fig. 6. The bed morphology for a
reduced width of B = 3.0 m and 2.0 m have a
uniform deposition rate over entire reservoir surface
and symmetric ripple patterns as shown in Fig. 5.
During the first hour of the tests, the observed flow
pattern was deviated to the right side. But, after two
hours the asymmetric jet was flipped from the right
side to the left one as shown in Fig.6. The footprints
of previous stages of the flow patterns are clearly
visible in Fig. 6(a). As a result of the asymmetry flip
flop from one side to the other, bed forms persisted
throughout the alternately deflected jet. An
asymmetric flow pattern has been observed for all
reduced width basins. Hence, the reservoir width did
not affect the asymmetric separation of the deflected
jet. However, the size of the center and the upstream
corner eddies were in accordance with the width.
- 136 -
Final bed deposition thickness [m]
Velocity magnitude [m/s]
]
0
0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
[
3000
0.020
0.018
0.015
3
0.018
2
0.018
0.
0.0250.025 035
0.020
0.020
0.025
0.0
30
0.025
0.020
5
03
0.
2000
0.020
8
0.01
0.015
1
0.018
0
(a)
2000
4000
6000
0.02 0.04
0.06 0.08
0.1direction
0.12 0.14
0.16 0.18 0.2
X-distance
in flow
[mm]
3
4000
6000
0.020
5
6
0.020
0
0.015
0.030
0.025
0.020
0.020
0.025
0.030
0.02
5
0.015
(d)
0
35
0.0
1
(c)
0.020
4
0.020
0.020
1000
2000
2
3
2
0
1
(b)
4
2000
0
0
0.000 0.010 0.015
0.020 0.025 0.030
0.035direction
0.040 0.045[m]
0.050 0.055 0.060
X-distance
in flow
0.025
3000
0
0.025
4000
0
B x L = 4.0 mx 3.0 m
0
0.030 0.020
0.020
1000
30
0.0
Y- distance in transversal
direction (mm)
[ ]
4
B x L = 4.0 mx 5.0 m
4000
0.010 0.015 0.018 0.020 0.025 0.030 0.035 0.040 0.045 0.050 0.055 0.06
0.020
1
2
3
4
5
6
X- distance in flow direction (mm)
Fig.8 Flow pattern with vectors and sediment deposition at 4.5 hrs of the reduced basin lengths L (a, b) 5.0 m, (c, d) 3.0 m.
By reducing the reservoir width, the decelerating
of the deviated jet is reduced. Final deposition
patterns were affected by the reservoir width, with
more symmetric and uniform distributions on the
entire surface, and concentrated deposits on both
sides. The evolution of bed at 1.5, 3.0, and 4.5 hrs at
the middle cross section of X = 3.0 m, for the
reduced widths of 3.0 and 2.0 m, are shown in Fig.7.
The depositions are progressively increased in the
downstream direction. The evolution process is
increasingly fast with speed of 0.66 cm/hrs, for both
widths. In fact, the velocity magnitude is increased
by reducing reservoir width. As a result, a fairly
deep scour hole develops in the bed profile near the
inlet as shown in Fig. 7. Proceeding with time, this
variation becomes slow due to the counteracting
effect of the jet flow and the formation of a
submerged channel under the centerline.
(3) Influence of the basin length
Influence of sediment transport
The average flow field and the corresponding
bed morphology contours after 4.5 hrs for the
reduced reservoir length are shown in Fig.8. In the
first ten minutes for a reduced basin length with
sediment entrainment, the flow is symmetric with
one circulation cell on each side. The reduction of
horizontal cells occurs with sediment transport. As
sediment is added to the flow, the turbulence is
reduced together with increasing roughness, which
causes an increase in velocity gradient when
compared to clear-water flow. The turbulence was
generated locally by the horizontal entrainment of a
mixture into the basin with stagnant water. The jet
pulse created a region of 3D turbulent flow,
characterized by mixing and entrainment. Then the
size of the turbulence increased rapidly. During the
subsequent stage of sediment settled down, the
horizontal motions are suppressed and the mixed
region becomes flat and the motion becomes quasi
vertical.
Figs. 8(b, d) show a high sediment deposition in
the center under the jet and low deposition near the
walls under the reversed jet. In both the experiments
ripples formed after about 3.0 hrs. By reducing the
basin length the flow is stabilized with a stable
symmetrical pattern. Four cells exist in the basin
length L = 5 m9). The four gyres interact with the jet,
which has some tendency to meander. By reducing
the basin length for L = 4, the number of gyres
remains constant and the flow pattern becomes
rather symmetric with respect to the centerline10).
The predominant change in the flow pattern is an
evolution from a four cells flow to a distinct two
cells flow. The length of the reservoir plays a
critical role in determining the jet flow type and the
associated bed deposition pattern. By reducing the
reservoir length as shown in Figs. 8 (a, c) the flow is
stabilized with a stable symmetrical pattern.
Asymmetric and switching flow behavior continues
towards the downstream end, where the jet is forced
to pass through the outlet channel.
(4) Drawdown flushing
The experiments in drawdown flushing were
started by opening the outlet gate and the water
depth reached to 0.1 m in the basin of the initial
water depth. Fig. 9 shows the flow structure and bed
thickness contours after two days of flushing in the
reduced basin width of 2.0 m. If the water surface
can be draw down significantly the flow starts to
erode a wide zone. At that stage, significant
amounts of sediments deposits were flushed through
- 137 -
1000
2
1
0.
01
8
0.018
0.020 0.015
0.030
0.0
0.030
28
20
0.0 0 0.022
. 02
0.030
0
22
0.0
0.030
5
03
0. 0.035
0.022
8 0.0
3
0.02
0.0 0
28
p
22
B = 2.0 m
2000
3
[ ]
0.0
3000
0.0000.0050.0100.0150.0200.0250.0300.0350.0400.0450.0500.0550.060
0 0.03 0.06 0.09 0.12 0.15 0.18 0.21 0.24 0.27 0.3
0
1
2
3
4
5
6
0
2000
4000
6000 0
Fig.9 Flow field and contours after two days flushing with lowering basin level to 50% (Q = 7.0 l/s and h = 0.1m)
0
the basin, and the initial flushing channel deepened
and widened as a result of the strong jet flow and
erosion. The flushing channel is defined as self
forming channel in the deposited material in the
basin and that will be maintained by the flushing jet
flow. Initially all deposited sediment resuspended
with the high velocity jet. The flushing channel
location was clearly visible by the re-suspended
sediment, which was rapidly released during 30
minutes. This followed by a short period of very
rapid widening of the flushing channel forming a
wide advancing front. Then, after a sudden change
in the rate of the channel widening, the flushing
channel continues to widen at a more gradual rate.
The larger basin volume will induce higher bed
shear stress, and hence produces more effective
removal of sediments on the bed. A significant
amount of sediment deposits was flushed through
the basin. To effectively apply the flushing
processes for removing deposits, the location, depth,
and width of the flushing channel can be changed by
modifications of the geometry or installation of
islands or a movable wall near the inlet jet. The
channel attracts the jet and stabilizes the flow
structures over the entire surface.
5. CONCLUSIONS
Although the geometry of a rectangular basin is
symmetric, the flow pattern is asymmetric under
certain conditions. The basin geometry influences
the behavior of the large turbulence structures, and
the flow is quite sensitive to the geometry shape.
When suspended sediment is added to the turbulent
flow over a plane bed the following was observed:
(a) the large coherent structures disappear compared
to clear water flow with similar flow properties and
(b) depositions and flow structure remains
asymmetric with reduced width of the reservoir and
disappears when reducing the length of the
reservoir. The findings can be used to define the
optimal shape of the stilling.
ACKNOWLEDGMENT: The authors gratefully
acknowledge the Swiss National Fund for Scientific
Research (SNFRS), which provided funding for the
research carried out during post-doctoral studies at
the Kyoto University.
REFERENCES
1) Kantoush, S.A.: Experimental study on the influence of the
geometry of shallow reservoirs on flow patterns and
sedimentation by suspended sediments, EPFL Thesis 4048
and Commun. (37) LCH, EPFL, ISSN 1661-1179, 2008.
2) Kantoush, S.A., De Cesare, G., Boillat, J.-L., and Schleiss, A.
J.: Flow field investigation in a rectangular shallow
reservoir using UVP, LSPIV and numerical modeling, J.
Flow Meas. and Instrum., Vol.19 (3-4), PP. 139-144, 2008.
3) Tung, Y.K. and Mays L.W.: Optimal Designs for Overflow
Spillway, Proceeding of ASCE, Journal of Hydr. Div.
ASCE, Vol. 108, No. 10, pp.1163-1178, 1982.
4) Rajaratnam, N., Subramanya, K.: Hydraulic jumps below
abrupt symmetrical expansions. J. Hydr. Div. ASCE, pp.
481-503, 1968.
5) Sumi, T. and Nakanishi, T.: Characteristics of Hydraulic
jump type energy dissipators below abrupt symmetrical
expansion, Civil Eng. Journal, 33-6, pp.21-27, 1991.
6) Bremen, R. and Hager, W. H.: T-jump in abrupt expanding
channel. J. Hydr. Res. 31(1), pp. 61–78, 1993.
7) Dewals, B. J., Kantoush, S. A., Erpicum, S., Pirotton, M., and
Schleiss, A. J.: Analysis of flow instabilities in shallow
rectangular basins, Envi. Fluid Mech., 8(31), 31–54, 2008.
8) Kantoush, S.A., Bollaert, E., and Schleiss, J.: Experimental
and numerical modeling of sedimentation in a rectangular
shallow basin, International Journal of Sediment Research,
Vol.23, No. 3, pp. 212–232, 2008.
9) Jia, Y. and Wang, S. S.Y.: CCHE2D: Two-dimensional
hydrodynamic and sediment transport model for unsteady
open channel flows over loose bed, Technical Report No.
NCCHE-TR-2001-1, National Center for Computational
Hydroscience and Engineering, The University of
Mississippi, USA.
10) Kantoush, S. A., and Schleiss, A. J.: Channel Formation in
Large Shallow Reservoirs with Different Geometries during
Flushing, J. Env. Technology, Vol.30, pp.1020-1032, 2009.
- 138 -
(Received September 30, 2009)